When someone asks me ‘why toxicology?’, I don’t think of the chance to use colorful antidotes like hydroxocobalamin for cyanide or Prussian blue for thallium toxicity. It’s not even the chance to manage a blue ringed octopus envenomation. Exotic cases are exciting, but for most of us, we choose toxicology because we are fueled by the excitement of sorting through a limited history and smattering of labs to discover a hidden disease process. We thrive on applying our understanding of mechanisms and historical evidence to pull a unifying diagnosis out of the obscure. Admittedly, though, the blue ringed octopus is extremely cool.
Toxic alcohols are a perfect example of the type of diagnostic challenge that drives us. Ethylene glycol (EG), methanol (MeOH), and isopropanol levels are almost universally unavailable in real time. The differential diagnosis for a toxic alcohol ingestion is huge, and although the alternative diagnoses are much more common, missing the diagnosis can sentence a patient to renal failure, permanent blindness, or death. Even more, toxic alcohol ingestions are accompanied by a great nemesis, a persistent villain that haunts nearly every consult: the osmol gap (OG).
But can’t we use the OG to make (or exclude) the diagnosis of toxic alcohols? Haven’t we been ordering this lab test for years, cognitively offloading any worry about tox as soon as the low OG results? Am I really going to ruin this for you? Yes, of course I am. Far from a quick fix solution, the OG remains one of the more misunderstood tests in medicine.
To understand this we must start with a few definitions:
Osmolarity & osmolality
Osmolarity is often used interchangeably with osmolality. The difference depends on whether what you measure is based on mass (osmolality measured as mOsm/kg) or volume (osmolarity measured as mOsm/L) . Mercifully, it doesn’t matter much because in serum the difference between the two is negligible. You can think of it as the number of independent particles in the serum. Osmolarity is primarily determined by the concentration of sodium cations and their counter ions, glucose, and blood urea nitrogen (BUN)1. Typical substances that alter osmolarity include sugars, alcohols and ketones.
Osmol gap (OG)
The osmol gap (OG) is the difference between the measured serum osmolality (a lab value you order that is measured by freezing point depression) and the calculated serum osmolarity. It indicates the number of unmeasured particles in serum, and in toxicology is designed to signal hidden poisons, specifically toxic alcohols. To calculate the serum osmolarity, you need a basic metabolic panel (BMP) and an ethanol level (see equation below). Due to the dynamic nature of metabolism, all components of the OG need to be drawn at the same time. For toxic alcohols, if you order the measured osmolality later than the BMP and ethanol level, a portion of the osmotically active parent compound will have metabolized and your measured osmolality will now be lower than they were when the BMP was drawn, reducing your ability to detect a gap.
With results from the BMP and EtOH level in hand, plug in values to the equation below, then subtract the results from your measured osmolality, and you have your OG.
(2Na) + (glucose/18) + (BUN/2.8) + (EtOH/4.6)
Where do we get this equation? It’s far from an exact science, and many equations have been proposed, all serving as approximations.2,3 Conversion factors based on molecular weight are used to get everything into compatible units.1
A large part of the difficulty in applying the OG stems from unreliability of the accepted normal values. This is in part due to these differing (and imperfect) equations, but also differing laboratory techniques, and population outliers from the mean used to calculate a normal gap. We run into serious trouble with the OG when we try to use it as a screening test. The classic teaching in medicine is that an OG 10 is sufficient to rule out a toxic alcohol ingestion (normal range -14 to +10)3, but applying this teaching to your patient can have deleterious and even life threatening consequences. The OG is simply not sensitive enough to be used as a “rule out” test.
To illustrate how a low OG can give false reassurance, imagine have a 3-year-old found with an open bottle of antifreeze. You get all the labs and calculate an OG of 3. Based on accepted values, this result is well within the perceived safe zone of 10. The problem is that you do not (and basically never will) know your patient’s baseline. Assuming she falls within the normal range (and in addition, 5% of the population will be further outliers), and has a baseline of -10, your true OG is now +13. In the case of an EG exposure, an OG of 13 translates to a concentration of 80 mg/dL (6.2 is the conversion factor for EG based on its molecular weight of 62 g/dL; 6.2 x 13 = 80.6). Clinically, we treat for EG levels over 20-25 mg/dL. Reliance on the OG as a “rule out” test just resulted in a missed toxic exposure in a child.
Although the above discussion should be the nail in the coffin of the OG as a useful screening test for a potentially lethal exposure, its failings continue. Parent alcohols produce a measurable OG, however their toxic metabolites do not. Over time, your patient clinically deteriorates as she metabolizes methanol into formic acid, but at the same time, the OG is simultaneously declining. See graph below. Although incredibly simple, it beautifully illustrates the time sensitive nature of the OG.
It is worth mentioning the volatile panel, which can be obtained at certain institutions (not many) and has a same-day turnaround time. Most people erroneously believe the volatile panel tests for all toxic alcohols, but EG is not volatile and is not included in the panel. Again, we are in a situation in which a negative test cannot rule out a toxic alcohol ingestion.
So what do we do if we can’t rely on a quick test or clear physical exam findings to make a potentially life threatening diagnosis? Here is the perspective of a medical toxicologist:
There are a few scenarios to consider.
First, the patient has a very high OG. And by high OG, we mean something greater than around 40. Although there are other causes of elevated osm gaps including alcoholic ketoacidosis (AKA), lactic acidosis, and renal failure, these conditions typically give more modest elevations in the osm gap. For osm gaps approaching or exceeding 40, antidotal therapy (fomepizole) should be initiated immediately, and toxicology should be consulted for the presumptive diagnosis of toxic alcohol ingestion.
A second scenario is one in which the clinical suspicion for toxic alcohol ingestion is very high. These are cases in which the patient has a history of previous toxic alcohol ingestion (yes, there are indeed frequent fliers in the world of toxic alcohols), or the patient presents saying “hi doc, i’m suicidal and i just drank antifreeze”. In these cases, our pretest probability is high and our threshold for initiating early treatment should be very low.
By far the most common scenario, however, is a patient who cannot provide a history and presents with a moderate AGMA and a non-diagnostic (i.e., 40) OG. There are many causes of AGMA and we’ve already shown low osm gaps to be relatively useless, so what should we do in these cases? To empirically treat all of these patients with fomepizole would quickly drain our national supply, incur unnecessary costs, and fill up our hospital beds.
Luckily, we do have something to go on to narrow the differential diagnosis, and that is evaluation of the AGMA. There are several mnemonics of varying levels of complexity and pain to “help” you sort through this differential, but I prefer to avoid those that contain medications i’ve never seen used in practice (looking at you MUDPILES) and to keep things simple. Other items on my differential diagnosis are ketosis (alcoholic ketoacidosis, starvation ketoacidosis, diabetic ketoacidosis), renal failure, lactate associated metabolic acidosis (metformin, ischemic gut, sepsis etc), and salicylates. There are a few other toxins that we may want to consider besides toxic alcohols and metformin (think CO, CN), but the history and physical would hopefully be more suggestive of these poisons. After CN exposure, for example, the patient typically presents on a spectrum of mostly dead to totally dead. If you like mnemonics, you’ll be happy to realize these all spell KULTS. Join the KULTS of toxicology. To evaluate this differential, order a beta-hydroxybutyrate if available and/or urine ketones, a lactate, and a salicylate level. You’ll pick up the uremia/renal failure on the BMP.
At the same time, or even before ordering these tests, initiate aggressive fluid resuscitation including thiamine and glucose. Thiamine is needed to get pyruvate into the Kreb’s cycle, and without it, pyruvate is converted to lactate, ATP production plummets and a metabolic acidosis ensues. The glucose will also help to resolve the ketotic alcoholic or starvation state. Reassess your patient with a repeat BMP, EtOH and measured serum osmolality after resuscitation. Fluids resuscitation alone won’t fix a methanol or ethylene glycol ingestion. Let’s say that again, if the AGMA resolves after fluids, the OG is not rising, and the EtOH is negative (remember EtOH, like fomepizole, blocks alcohol dehydrogenase and prevents metabolism of toxic alcohols to respective toxic metabolites) you’ve pretty much ruled out a significant toxic alcohol ingestion. If the patient remains with a refractory AGMA, and there is no alternative diagnosis from your KULTS work up, treat the patient with fomepizole and consult toxicology for presumed diagnosis of a toxic alcohol.
There are a few other tricks to help you with the diagnosis of a toxic alcohol. My favorite is the “lactate gap” for EG cases. Ethylene glycol, like ethanol, should not significantly raise your serum lactate (mild elevations of 4-5 can be seen due to unfavorable NADH/NAD ratios shifting pyruvate to lactate). Many (but not all) blood gas analyzers, however, will read very high lactates due to cross reactivity of the EG metabolite glycolate with L-lactate oxidase.4 A high blood gas lactate with a low serum lactate should significantly raise suspicion for EG in the right clinical setting. Lastly, single digit bicarbs should give you pause. Other pathologies on this differential can drive the bicarb down to low levels, but when you see that single digit bicarb, make sure you can convince yourself of the alternative diagnosis before moving on.
I’ve avoided discussing crystalluria or urine fluorescence for EG, as these tests are neither sensitive nor specific and have basically no role in the diagnostic process. Acute kidney injury (remember those calcium oxalate crystals, they can also deposit in the heart and cranial nerves!) and vision loss can be seen in EG and MeOH toxicity respectively. These are both late findings, however, and if you are relying on them, you and your patient are in trouble.
So, there you have it: the diagnosis of a toxic alcohol ingestion from the mind of a toxicologist. And if you are inspired by the destruction of nemeses, MacGyvering of diagnoses, and pulling the Kreb’s cycle into every discussion, maybe toxicology is for you. Plus we include gratuitous videos of adorable venomous animals. Come join the KULT.